Many operations and maintenance (O&M) professionals have encountered scenarios where timestamps in server logs differ from actual business event times by over ten seconds—making troubleshooting impossible—or where clock desynchronization between nodes in a distributed system leads to data consistency issues. The root cause of these problems often points to a single issue: improper NTP (Network Time Protocol) configuration.

NTP is one of the oldest and most fundamental protocols on the Internet. It utilizes a hierarchical time-source architecture to maintain millisecond-level time synchronization across hundreds of millions of devices worldwide. For servers, time synchronization is not merely an optional enhancement; it is a prerequisite for log auditing, security authentication, and distributed system coordination.

This article outlines the categories, selection criteria, and configuration methods for common NTP servers, helping you identify the time source best suited to your specific business needs.

Based on the operating entity and use case, commonly used NTP servers in China generally fall into four categories: authoritative national time sources, public NTP services provided by cloud vendors, time services from education and research networks, and the globally distributed NTP Pool.

Key Criteria for Selecting NTP Servers

Now that the available server types are known, the next question is: how do you choose?

Latency is the most direct factor. NTP is highly sensitive to network jitter. Round-trip time (RTT) for international links typically exceeds 200ms and is significantly affected by fluctuations in international gateway bandwidth, whereas RTT between domestic nodes is usually kept within the 10ms to 50ms range. Low latency translates to higher synchronization precision, allowing server clock errors to be kept within the millisecond range. Therefore, prioritizing servers that are physically closer is a fundamental principle.

Reliability depends on hierarchy and redundancy. The NTP protocol uses "Stratum" levels to measure the reliability of a time source; a lower Stratum indicates proximity to the primary time source and higher precision. However, relying on a single time source always carries a risk of failure. Nutanix configuration guidelines explicitly recommend configuring at least five stable, reliable time sources. Three is the absolute minimum, but this only allows for the detection of a "false clock" without providing redundancy; four is the minimum standard for redundancy; and five is the recommended number.

Security is becoming increasingly critical as we approach 2026. The NTP protocol itself lacks built-in encryption mechanisms, making it vulnerable to NTP reflection attacks and man-in-the-middle tampering. For cross-border e-commerce payment systems or operations involving sensitive data, it is advisable to use servers that support NTS (Network Time Security) or to protect the time synchronization process using NTP authentication mechanisms. Additionally, avoid using rate-limited NTP servers; such servers may drop packets during periods of high concurrent requests, leading to synchronization interruptions.

Address stability is another important factor. It is recommended to use an FQDN (Fully Qualified Domain Name) rather than an IP address when specifying public NTP servers. IP addresses may change due to network adjustments by the service provider; using a domain name allows the system to automatically adapt to DNS updates, thereby minimizing synchronization interruptions caused by IP changes.

Recommendations for Different Scenarios

Individual developers and small projects: Selecting public NTP addresses from one or two cloud providers is sufficient—for example, using `ntp.aliyun.com` as the primary source and `ntp.tencent.com` as a backup. These addresses offer broad coverage, high stability, and simple, out-of-the-box configuration.

Enterprise production environments: It is recommended to configure three to five time sources of varying types. You can combine sources such as the National Time Service Center, public NTP servers from cloud providers, and NTP Pool nodes. Users of domestic VPS services should also consider the server's physical location; if a server is deployed in Hong Kong or overseas, selecting a geographically closer NTP server (such as `time.google.com` or a node provided by a local ISP) ensures better synchronization quality than forcing a connection to a domestic source. Technical documentation for Jtti's overseas cloud servers also advises manually specifying geographically closer time sources to reduce synchronization latency.

Large-scale clusters and distributed systems: The best practice is to set up a dedicated NTP server within the cluster to serve as a unified reference point, while calibrating it against a reliable upstream time source. This approach reduces concurrent load on external time sources while ensuring perfect time consistency across all nodes within the cluster.

Configuration Considerations and Common Issues

Regardless of whether you use NTP or Chrony, the core configuration logic remains the same. After installing Chrony on Ubuntu or Debian systems, edit `/etc/chrony.conf` and add time source addresses at the top of the file. Chrony outperforms traditional NTP in virtualized and unstable network environments, making it particularly suitable for overseas cloud servers.

server ntp.aliyun.com iburst

server ntp.tencent.com iburst

server ntp.ntsc.ac.cn iburst

The `iburst` parameter allows the system to quickly send a series of packets upon startup, accelerating the initial synchronization. Once configured, restart the service using `systemctl restart chronyd` and check the synchronization status with `chronyc sources`; sources marked with a specific symbol (usually `*`) in the output indicate they are currently in use.

On Windows systems, you can manually enter NTP server addresses for synchronization via "Control Panel > Date and Time > Internet Time."

Be aware of a few common issues during configuration. You must allow traffic on UDP port 123, the default port for NTP communication. Pay special attention to time drift in virtualized environments; time drift on the host machine of some overseas cloud providers can propagate to the virtual machine. In such cases, disable the virtual machine's automatic RTC synchronization and rely instead on NTP or Chrony for network-based synchronization. Regularly checking the synchronization status is also crucial; commands like `ntpq -p` or `chronyc tracking` allow you to monitor offset and latency. A consistently large offset indicates that the time source needs to be changed.

Time Synchronization and Server Selection

While NTP server configuration addresses time synchronization at the software layer, the ultimate precision and stability of synchronization depend on underlying network latency and jitter. If the server's network connection is unstable, synchronization requests may frequently fail due to packet loss and latency fluctuations, regardless of how high-quality the time source is.

Jtti's cloud server solutions provide the infrastructure to support this effectively. Their Hong Kong and US nodes utilize premium CN2 GIA lines with optimized direct connectivity across major carriers, ensuring low latency and minimal jitter for access from within China. This results in more predictable response times for NTP synchronization requests when managing overseas servers from a domestic location. Dedicated bandwidth comes standard across the entire product line, eliminating issues where "noisy neighbors" saturate bandwidth and cause synchronization request timeouts. For long-running applications requiring precise timing, a lifetime recurring discount policy ensures server costs remain stable—even amidst market price hikes—providing cost predictability for long-term time synchronization operations.

There is no single "best" NTP server; there is only the one that is most suitable for your needs. Prioritize low-latency nodes from providers like Alibaba Cloud or Tencent Cloud for domestic operations, and select time sources with closer physical proximity for overseas servers. For security-sensitive scenarios, consider servers that support NTS. Regardless of the use case, configure at least three time sources to ensure redundancy. Visit the official Jtti website to view full specifications and test IPs for Hong Kong CN2 and US CN2 cloud servers, and choose a solid network foundation for your time synchronization solution.